The invention relates to a method for detecting the beginning of the dips of the amplitudes in the output signal of a receiver for time signals identifying the beginning of a second, where the output signal of the receiver for time signals is sampled N-times per second; the sampled values are stored in a memory field with N cells; in the memory field a mean signal curve for the time interval of one second is generated over a period of several seconds; and the beginning of the seconds is determined from the mean signal curve.
A method for recovering information from the faulty data of a time-signal transmitter is known from DE 37 33 965 A1, where a signal obtained from the signal curves over several seconds for the purpose of synchronizing the radio-controlled clock to the beginning of the second is compared with a model signal. In this known method, the signal supplied from the receiver is sampled with a preset frequency. In a time interval of one second, the sampled values are cumulated at the corresponding time positions. After a certain time, a mean signal curve is formed. In each of the two types of signal that occur in the time-signal transmitter DCF-77, the falling edge is the same. To determine the beginning of the second, a correlation is applied between the mean signal obtained from several seconds curves and a model signal.
This known method has the considerable disadvantage of comparing with a model signal. Considerable additional effort is required to provide the model signal in the form of a table or a computational rule. A great deal of computation is needed in addition in order to perform the comparison using correlation methods.